To avoid service interruption, some key service processing devices in a network have backup devices. In addition, to meet a reliability requirement of the network, the backup devices usually need to be deployed remotely. Therefore, in the network, a disaster tolerance solution for the devices deployed remotely is required.
Currently, a dual-host disaster tolerance technology may adopt a virtual route redundancy protocol (VRRP) solution, in which a virtual router needs to be created in the network, and the virtual router is composed of an active router and several standby routers. Both the active router and the standby routers need to receive an address resolution protocol (ARP) request packet sent by a host, and then, the active router makes a response. In this solution, the active router and the standby routers are required to be located in a same layer 2 switching network. However, the devices deployed remotely usually interwork in a layer 3 switching network and a remote layer 2 switching network is difficult to be provided, so the VRRP solution is not suitable for remote disaster tolerance. In addition, for the remote disaster tolerance, a manner of configuring different IP addresses for an active device and standby devices can be adopted. This manner needs to configure, on a host, IP addresses corresponding to the active device and the standby devices, and after the active device is abnormal, an IP address of a standby device is automatically switched to and used. Therefore, this solution has a capacity requirement for the device, and after the IP address is changed, a connection needs to be reestablished, which may cause service interruption.